\begin {table}[ht]\centering
  \begin{threeparttable}
\setlength\extrarowheight{-3pt}
  \caption{Probability Republican Respondent Approves of President Defining Republicans By Baseline or Survey Response}
\begin{tabular}{lcccc}
\hline\hline
                    &\multicolumn{1}{c}{(1)}         &\multicolumn{1}{c}{(2)}         &\multicolumn{1}{c}{(3)}         &\multicolumn{1}{c}{(4)}         \\
                    &                     &                     &                     &                     \\
\hline
Republican In Survey&      $$estout_1$$ &   $$estout_2$$  &     $$estout_3$$ &    $$estout_4$$ \\
                    &   ($$p_1$$)         &    ($$p_2$$)         &   ($$p_3$$)         &   ($$p_4$$)         \\
[1em]
\hline
Observations        &        $$N_1$$         &        $$N_2$$         &       $$N_3$$         &        $$N_4$$         \\
\(R^{2}\)           &       $$R2_1$$         &       $$R2_4$$         &       $$R2_3$$         &       $$R2_4$$         \\
Weighted            &          No         &         Yes         &          No         &         Yes         \\
Covariates          &          No         &          No         &         Yes         &         Yes         \\
[1em]
\hline
Mean Approval Among Baseline Republicans &        $$mean_baseline_1$$         &        $$mean_baseline_2$$         &       $$mean_baseline_3$$         &        $$mean_baseline_4$$         \\ 
[1em]
 \hline
\end{tabular}
$^{+} p<0.10,  ^{*} p<0.05, ^{**} p<0.01,  ^{***} p<0.001$
\label{AppOutcomeRepBaselineSurvey}
   
       \begin{tablenotes}
      \small
      \item The table reports results of a regression of approval on an indicator for whether the respondent identified as a Republican in the survey (as opposed to the excluded group of identifying as a Republican in the baseline period). The sample is constructed by including only respondents who identify as a Republican in either the survey or the baseline period and stacking the observations. An individual respondent would appear in the sample twice if the respondent self-reported as a Republican in both the baseline and the survey. Two-sided p-values calculated from heteroskedasticity robust standard errors clustered at respondent level are reported in parentheses.
      \item  

    \end{tablenotes}
  \end{threeparttable}
\end{table}